Global completion of the AKSZ classification of QP-structures on the parity-reversed tangent bundle $\Pi TN$ (geometry of the master equation)

The Batalin-Vilkovisky (BV) master equation, the consistency condition of the standard covariant quantization scheme for gauge theories, was given a geometric formulation by Alexandrov, Kontsevich, Schwarz and Zaboronsky (AKSZ) in 1997: its classical solutions live on QP-manifolds, supermanifolds equipped with an odd symplectic form invariant under an odd nilpotent vector field $Q$. AKSZ proved a classification theorem for such structures on the parity-reversed tangent bundle $\Pi TN$ - the target geometry of their topological sigma-model construction - but only in a conditional form: assuming $N$ even and simply connected, working near the zero section or within small deformations, and passing over an integration step that the authors themselves flagged as incomplete ('Probably one can give a complete proof of the statements a), b) in this way'). No later work has completed, globalized, or refuted this classification: adjacent results either establish local normal forms for Z-graded Q-manifolds, treat the master equation on affine varieties up to a coarser stable equivalence, transpose the question to derived stacks with shifted symplectic structures, or remain deliberately formal. The open problem is to complete the AKSZ classification into an unconditional global theorem: a gap-free proof of statements a) and b), removal of the simply-connectedness, parity, and smallness hypotheses, and explicit verifiable conditions under which infinitesimal $Q$-preserving canonical transformations integrate to finite diffeomorphisms homotopic to the identity.

Background

The Batalin-Vilkovisky (BV) formalism is the standard covariant quantization scheme for gauge theories: it places fields, ghosts (fields of reversed statistics encoding the gauge symmetries) and antifields on an equal footing on a supermanifold - a space with both commuting and anticommuting coordinates - equipped with an odd Poisson bracket, the antibracket $\{\cdot,\cdot\}$ (a bilinear operation obeying graded analogues of antisymmetry and the Jacobi identity that raises Grassmann parity). The consistency condition of the scheme is the classical master equation $\{S,S\}=0$ for the extended action $S$. Schwarz's geometric reformulation ('Geometry of Batalin-Vilkovisky quantization') identifies the arena as a P-manifold, a supermanifold with an odd symplectic form (a closed, non-degenerate two-form of odd Grassmann parity), with an SP-manifold adding a compatible volume element for the quantum master equation.

The 1997 paper of Alexandrov, Kontsevich, Schwarz and Zaboronsky (AKSZ) recast this geometry and connected it to topological field theory. Its basic objects are: a Q-manifold, a supermanifold carrying an odd vector field $Q$ with $Q^2=0$ (a homological vector field, so called because it makes functions into a cochain complex); a P-manifold as above; and a QP-manifold, a supermanifold that is both, with the odd symplectic form $\omega$ invariant under $Q$ ($\mathcal{L}_Q\omega=0$). The central example is the parity-reversed tangent bundle $\Pi TN$, obtained from the tangent bundle of a smooth manifold $N$ by reversing the Grassmann parity of the fiber coordinates, written $(x,\eta)$ with $\eta$ the odd partner of a tangent vector; the de Rham differential of $N$, expressed as an odd vector field, is the standard homological field on $\Pi TN$. The Taylor coefficients of a general homological field at a point organize into an $L_\infty$ algebra (a strongly homotopy Lie algebra: a graded vector space with multilinear brackets satisfying Jacobi identities up to coherent higher homotopies), and AKSZ proved this correspondence reversible in a formal neighborhood - whether it holds on a genuine (convergent) neighborhood is one of the issues below.

The paper's main geometric result (its statements a) and b)) classifies QP-structures on $\Pi TN$ with the standard $Q$, under the standing hypotheses that $N$ is an even, simply connected manifold and that the pointwise homology groups $H_m=\ker Q_m/\operatorname{im}Q_m$ of the linearization of $Q$ at points $m$ of the zero locus $Y$ of $Q$ (the set where $Q$ vanishes) are trivial - a condition that holds automatically for $\Pi TN$ itself, is stable under small deformations (a small deformation can only decrease these groups), and is what allows a general QP-manifold to be reduced, near $Y$, to $\Pi TY$ with the standard field. The classification states that such structures correspond one-to-one to Omega-forms: odd symplectic forms built from a finite collection $\Omega=\sum_n\Omega_n$ of closed differential forms on $N$, with $\Omega_2$ symplectic, via $\omega=d(\partial\Omega/\partial\eta\, dx)$; and that two of them are equivalent - related by a diffeomorphism carrying one odd symplectic form to the other, preserving $Q$, and homotopic to the identity (a $Q$-preserving canonical transformation) - exactly when $\Omega'_n-\Omega_n$ is exact for every $n$ and $\Omega_2$ and $\Omega'_2$ determine equivalent symplectic structures on $N$, that is, are connected by a continuous family of non-degenerate two-forms all lying in the same de Rham cohomology class. The proof first performs this reduction, and then integrates infinitesimal canonical transformations to finite ones; these transformations act through two functions, $\sigma$ (entering the normal form $\omega=d(\partial\Omega/\partial\eta\, dx)+\mathcal{L}_Qd\sigma$) and $F$ (realizing the infinitesimal cohomological shift $\Omega'-\Omega=\hat{Q}F$, where $\hat{Q}$ denotes the action of $Q$ on functions).

This theorem is the target-space backbone of the AKSZ construction of topological sigma models (two-dimensional topological field theories, including the A-model and B-model obtained by twisting supersymmetric sigma models, and many descendants such as the Poisson sigma model (a two-dimensional topological sigma model whose target is a Poisson manifold, i.e. a manifold whose functions carry a bracket satisfying the Jacobi identity and the Leibniz rule)): the classical BV phase space of such a model is a mapping space into a QP-manifold, so the classification decides exactly which target geometries - hence which topological field theories - the construction delivers; deformations of the target QP-structure are labeled by the de Rham cohomology of $N$.

The proof is, however, explicitly conditional at three flagged points. First, the hypotheses: $N$ is assumed even and simply connected, results for a supermanifold $N$ are stated only in a small neighborhood of $N$, and the class covered is that of trivial $H_m$ - in effect, structures obtained from $\Pi TN$ by small deformations. Second, the passage from infinitesimal to finite canonical transformations is asserted rather than proven: the authors write that conditions on the generator $\sigma$ 'can be imposed without loss of generality', and, of an alternative proof route they sketch, 'Probably one can give a complete proof of the statements a), b) in this way.' Third, the $L_\infty$ data obtained from the Taylor series of $Q$ yields a genuine homological field only where that series converges; otherwise the structure lives only on a formal ('infinitesimal') neighborhood. Later literature has built a large theory of Q-manifolds and AKSZ-type models, but no work completes, globalizes, or refutes the classification itself; the 2026 review of the subject (Ikeda) still describes the QP-geometry of mapping spaces as not well understood. The open problem is to turn the AKSZ classification into an unconditional global theorem - or to show by explicit example that it cannot be.

Problem Statement

Complete - or refute - the AKSZ classification of QP-structures on $\Pi TN$, turning the conditional, local, small-deformation statements a) and b) of the source paper into an unconditional global theorem.

Setup. Let $N$ be a smooth manifold, $\Pi TN$ its parity-reversed tangent bundle with coordinates $(x,\eta)$ ($\eta$ the odd partner of a tangent vector) and the standard (de Rham) homological vector field $Q$ (odd, with $Q^2=0$). Let $H_m=\ker Q_m/\operatorname{im}Q_m$ denote the homology of the linearization of $Q$ at a point $m$ of the zero locus $Y$ of $Q$ (the set where $Q$ vanishes); for the standard $Q$ on $\Pi TN$ these groups are automatically trivial, and their triviality is the hypothesis under which a general QP-manifold reduces, near the zero locus of its own $Q$, to $\Pi TY$ with the standard field. Two QP-structures on $\Pi TN$ are equivalent if they are related by a $Q$-preserving canonical transformation (a diffeomorphism carrying one odd symplectic form to the other and preserving $Q$) homotopic to the identity. An Omega-form is the odd symplectic form $\omega=d(\partial\Omega/\partial\eta\, dx)$ determined by a finite collection $\Omega=\sum_n\Omega_n$ of closed forms on $N$ with $\Omega_2$ symplectic.

(i) Give a complete, gap-free proof of the correspondence: QP-structures on $\Pi TN$ (in the widest class for which it holds, at minimum the class with trivial $H_m$) correspond one-to-one to Omega-forms, two structures being equivalent exactly when $\Omega'_n-\Omega_n$ is exact for every $n$ and $\Omega_2$ and $\Omega'_2$ determine equivalent symplectic structures on $N$, that is, are connected by a continuous family of non-degenerate two-forms all lying in the same de Rham cohomology class. In particular, the step integrating infinitesimal $Q$-preserving canonical transformations to finite ones must be proven, not assumed.

(ii) Remove the standing hypotheses and state what survives: $N$ even; $N$ simply connected; the restriction to neighborhoods (of the zero section, or of $N$ when $N$ is a supermanifold); and the smallness of deformations. State precisely for which $N$ the correspondence with Omega-forms holds, and what data (for example monodromy of the class $[\Omega]$) replaces it in general.

(iii) Give explicit, verifiable conditions on the data $(\sigma,F)$ through which an infinitesimal $Q$-preserving canonical transformation acts - $\sigma$ the function in the normal form $\omega=d(\partial\Omega/\partial\eta\, dx)+\mathcal{L}_Qd\sigma$, $F$ the function realizing the infinitesimal shift $\Omega'-\Omega=\hat{Q}F$ of the Omega-data ($\hat{Q}$ the action of $Q$ on functions) - under which the transformation integrates to a finite $Q$-preserving canonical diffeomorphism homotopic to the identity; prove the conditions sufficient and, if possible, necessary.

The classification in (i) fails if there exists an explicit QP-structure on $\Pi TN$ in the class over which the correspondence is claimed (at minimum the class with trivial $H_m$) that is not equivalent to any Omega-form structure, or two inequivalent QP-structures in that class with the same Omega-data; a structure outside that class only delimits the widest class, the subject of (ii).

References

1. M. Alexandrov, M. Kontsevich, A. Schwarz, O. Zaboronsky, 'The geometry of the master equation and topological quantum field theory', International Journal of Modern Physics A 12 (1997) 1405-1429, arXiv:hep-th/9502010, DOI 10.1142/S0217751X97001031, https://arxiv.org/abs/hep-th/9502010
2. A. Schwarz, 'Geometry of Batalin-Vilkovisky quantization', Communications in Mathematical Physics 155 (1993) 249-260, DOI 10.1007/BF02097392, arXiv:hep-th/9205088, https://arxiv.org/abs/hep-th/9205088
3. H. Khudaverdian, 'Semidensities on odd symplectic supermanifolds', Communications in Mathematical Physics 247 (2004) 353-390, DOI 10.1007/s00220-004-1083-x, arXiv:math/0012256, https://arxiv.org/abs/math/0012256
4. H. M. Khudaverdian, 'Laplacians in odd symplectic geometry', Contemporary Mathematics 315 (2002) 199-212, DOI 10.1090/conm/315/05481, arXiv:math/0212354, https://arxiv.org/abs/math/0212354
5. H. M. Khudaverdian, M. Peddie, 'Odd Laplacians: geometrical meaning of potential, and modular class', Letters in Mathematical Physics 107 (2017) 1195-1214, DOI 10.1007/s11005-017-0938-y, arXiv:1509.05686, https://arxiv.org/abs/1509.05686
6. P. Ševera, M. Širaň, 'Integration of differential graded manifolds', International Mathematics Research Notices 2020 (2020) 6769-6814, DOI 10.1093/imrn/rnz004, arXiv:1506.04898, https://arxiv.org/abs/1506.04898
7. A. Kotov, C. Laurent-Gengoux, V. Salnikov, 'Normal forms of $\mathbb{Z}$-graded $Q$-manifolds', Journal of Geometry and Physics 191 (2023) 104908, DOI 10.1016/j.geomphys.2023.104908, arXiv:2212.05579, https://arxiv.org/abs/2212.05579
8. S. Seol, M. Stiénon, P. Xu, 'Dg manifolds, formal exponential maps and homotopy Lie algebras', Communications in Mathematical Physics 391 (2022) 33-76, DOI 10.1007/s00220-021-04265-x, arXiv:2106.00812, https://arxiv.org/abs/2106.00812
9. A. S. Arvanitakis, 'Formal exponentials and linearisations of QP-manifolds', arXiv:2204.12613, https://arxiv.org/abs/2204.12613
10. Th. Th. Voronov, 'Vector fields on mapping spaces and a converse to the AKSZ construction', arXiv:1211.6319, https://arxiv.org/abs/1211.6319
11. A. S. Arvanitakis, D. Tennyson, 'Brane wrapping, AKSZ sigma models, and QP manifolds', arXiv:2301.02670, https://arxiv.org/abs/2301.02670
12. G. Felder, D. Kazhdan, 'The classical master equation', arXiv:1212.1631, https://arxiv.org/abs/1212.1631
13. S. L. Lyakhovich, E. A. Mosman, A. A. Sharapov, 'Characteristic classes of Q-manifolds: classification and applications', Journal of Geometry and Physics 60 (2010) 729-759, DOI 10.1016/j.geomphys.2010.01.008, arXiv:0906.0466, https://arxiv.org/abs/0906.0466
14. S. L. Lyakhovich, E. A. Mosman, A. A. Sharapov, 'On characteristic classes of Q-manifolds', arXiv:math/0612579, https://arxiv.org/abs/math/0612579
15. M. A. Grigoriev, A. M. Semikhatov, I. Yu. Tipunin, 'Becchi-Rouet-Stora-Tyutin formalism and zero locus reduction', Journal of Mathematical Physics 42 (2001) 3315-3333, DOI 10.1063/1.1367867, arXiv:hep-th/0001081, https://arxiv.org/abs/hep-th/0001081
16. Si Li, 'Vertex algebras and quantum master equation', Journal of Differential Geometry 123 (2023), DOI 10.4310/jdg/1683307007, arXiv:1612.01292, https://arxiv.org/abs/1612.01292
17. Z. Kökényesi, A. Sinkovics, R. J. Szabo, 'Double field theory for the A/B-models and topological S-duality in generalized geometry', Fortschritte der Physik 66 (2018) 1800069, DOI 10.1002/prop.201800069, arXiv:1805.11485, https://arxiv.org/abs/1805.11485
18. S. Noja, 'On BV supermanifolds and the super Atiyah class', European Journal of Mathematics 9 (2023), DOI 10.1007/s40879-023-00603-6, arXiv:2202.08136, https://arxiv.org/abs/2202.08136
19. N. Ikeda, 'Q-Manifolds and Sigma Models', in Geometric Methods in Physics XLI (Białystok, 2024), Trends in Mathematics, Springer (2025), arXiv:2604.23496, https://arxiv.org/abs/2604.23496
20. Th. Th. Voronov, 'Graded Geometry, Q-Manifolds, and Microformal Geometry', Fortschritte der Physik 67 (2019) 1910023, DOI 10.1002/prop.201910023, arXiv:1903.02884, https://arxiv.org/abs/1903.02884
21. C. Arias, N. Boulanger, P. Sundell, A. Torres-Gomez, '2D sigma models and differential Poisson algebras', Journal of High Energy Physics 2015 (2015) 095, DOI 10.1007/JHEP08(2015)095, arXiv:1503.05625, https://arxiv.org/abs/1503.05625
22. C. Arias, P. Sundell, 'A-type Sigma Models from Differential Poisson Geometry', arXiv:2607.29668, https://arxiv.org/abs/2607.29668
23. T. Pantev, B. Toën, M. Vaquié, G. Vezzosi, 'Shifted symplectic structures', Publications Mathématiques de l'IHÉS 117 (2013) 271-328, DOI 10.1007/s10240-013-0054-1, arXiv:1111.3209, https://arxiv.org/abs/1111.3209
24. E. Bouaziz, I. Grojnowski, 'A d-shifted Darboux theorem', arXiv:1309.2197, https://arxiv.org/abs/1309.2197